fix(pipeline): premature state update + reject 재생성 알림

버그1: /feedback approve가 bg task 시작 전에 state를 next_pending으로 set →
polling이 빈 video_url로 알림 발송. bg task의 run_step이 state를 set하도록
일임 — 이중 update 제거.

버그2: reject 후 같은 *_pending 상태로 재생성됐을 때 dedupe에 막혀 알림이
안 감. dedupe 키에 feedback_count_per_step[step]을 포함 — 재생성마다
count가 증가하므로 키가 달라져 재알림 동작.
This commit is contained in:
2026-05-08 23:08:24 +09:00
parent 6d416aab78
commit 97b15cb985
3 changed files with 40 additions and 6 deletions

View File

@@ -1009,11 +1009,19 @@ async def feedback(pid: int, req: FeedbackRequest, bg: BackgroundTasks):
if req.intent == "approve":
from .pipeline.state_machine import next_state_on_approve
next_st = next_state_on_approve(state)
_db_module.update_pipeline_state(pid, next_st)
# Validate transition is legal
try:
next_st = next_state_on_approve(state)
except ValueError as e:
raise HTTPException(400, str(e))
next_step = _state_to_step(next_st)
if next_step:
# bg task will set state to the new *_pending when step completes
bg.add_task(orchestrator.run_step, pid, next_step)
else:
# No step to run — fall through to direct state update
# (defensive — current code paths don't hit this)
_db_module.update_pipeline_state(pid, next_st)
return {"ok": True}
elif req.intent == "reject":